Last updated on October 8, 2014
My paper with my former colleague Nick Carnes on the political economy of agricultural protection, which looks at why members of Congress vote for or against the farm bill, has just been accepted by and is now forthcoming at Food Policy.
Nick and I have been working on this since early summer 2013. The idea came to us as we were having lunch and I asked him whether the data he’d assembled for his book White Collar Government included any information about whether members of Congress had worked as farmers. We began with an interest in knowing whether having worked as a farmer drove how members of Congress vote on farm bills, but we soon expanded our analysis to also include the proportion of farmers in a member of Congress’ district as well as the amount of money she had received from agricultural political action committees. This allows us to run a horse race between competing theories for what drives agricultural protection, i.e., subsidies to farmers along with the taxes and quotas imposed on agricultural imports.
After the paper was rejected twice at other journals (complete with a referee report that compared part of our paper to a high-school term paper, no less) we decided to send this to Food Policy in October 2013. Little did I know that I was going to become an associate editor of that same journal a month later!
Here is the abstract:
It seems paradoxical that until recently, developed countries have continued subsidizing agriculture even though their agricultural sectors had been declining in relative importance since the middle of the 20th century. What drives support for agricultural protection—the broad array of subsidies to farmers and taxes and quotas imposed on agricultural imports—in developed countries? We answer this question by testing three competing hypotheses about what drives support for agricultural protection in the US: (i) legislator preferences, (ii) electoral incentives, or (iii) lobbying. Using data on the roll call votes of the members of the 106th through the 110th Congresses (1999-2009) and the scores given to each legislator by the Farm Bureau, our findings suggest that electoral incentives explain a great deal of the variation in support for agricultural protection, but that legislator preferences and lobbying might play a role, too. Moreover, legislator preferences and electoral incentives appear to be substitutes for one another. Why does Congress support agricultural protection? Because many members have electoral incentives to—and because many of those who do not still have other personal or strategic interests at stake.
You can download the accepted version of our paper here. The analysis does not establish causal relationships, but it is interesting in the sense that, to me at least, it clears up the common misconception that lobbying drives everything. What we find instead is that it looks as though the electorate mainly drives whether farm bills get passed.
Marc: I was very excited to read your paper but there are some part of it that trouble me. It seems a very good first-cut, but here are a few concerns. I’d be interested to know your thoughts/reactions:
1) Unless you control for legislators who have no intention of running for re-election, the PAC contributions are still entirely endogenous–as you argue they should be in the paper. The moral hazard problem (taking the money and voting against) only exists if it’s not a repeated game.
2) There is a self-selection problem with ag committee membership. You control for being on the committee, but the fact they are on the committee reflects the attributes you’re trying to test for. This is seems to be reflected in your results. The Senate was effectively controlled by Democrats in both periods, giving them a majority on the ag committee–so a greater probability that any one committee member is Democratic/anti-farm supports, which is borne out in your results. The House (and therefore ag committee) was controlled by Republicans in 2002 and Democrats in 2008. In 2002, House Committee members (on average Republicans) voted for the bill while in 2008 the results are much weaker (in magnitude and significance).
3) It’s difficult to argue that even 2% of the population of a district being in any one industry could meaningfully affect the outcome of an election (i.e., make an incumbent lose sleep about reelection) unless the election is very tight. Even more so if their interests are in fact diametrically opposed to the interests of the other 98%. Did you consider using the percent of an electoral district’s economy that is ag-based, as opposed to the percentage of farm owners/mangers? One would imagine that electoral districts in which there are more farm equipment dealers, more elevators, even more farm equipment manufacturers, would have even more meaningful constituency support for the farm bill. Percentage of ‘rural’ square miles might also be a more meaningful proxy, since the tax base of those regions tend to be more dependent on the ag sector.
Finally, I am confused by your use of the Farm Bureau information. I expected that you were going to use that at some point as an instrument within the electoral equation, in a two-stage modeling system perhaps. I’m also curious what happens if you just use a dichotomous variable for farm experience rather than trying to capture a percentage of career. I’m sure why percentage would matter, unless the specific career path would have created an off-setting personal interest–especially if a large portion of career was spent in Congress.
The logrolling issue on SNAP/NSLP vs farm aid is obviously an issue and one of the reasons I was looking forward to read the paper. The way the 2014 Farm Bill played out (with threats to divorce food and farm assistance programs) illustrates how much tension there is that historical quid pro quo. I’m curious how much your results reflect how that tension has been building over the past legislative cycles.
Mike,
Thanks for those detailed comments. It’s a shame you weren’t one of the reviewers; it would have made for a much stronger piece. Thank you so much for reading and for commenting. Nick and I both appreciate it a great deal, and FWIW, the data will be made available soon if you’d like to have a go at them.
1. Agreed–in part. I think a lot of those legislators don’t renege in the last period like one would expect them to, perhaps because they don’t want to burn bridges for their post-Congress career. But that is indeed an empirical question, and we should have controlled for this.
2. Absolutely. I’m not sure this isn’t controlled for by observables, however.
3. We argue in the paper that our measure of electoral competition is a proxy for how many people depend on agriculture in a district. The 2% figure is low indeed, but think of how many other people depend on those 2%, i.e., workers, providers of services, dealers of seeds, fertilizer, pesticides, implements, etc. so that we are confident that this is a good proxy for how many people in a district have a stake in the farm bill getting passed.
4. I don’t think Farm Bureau scores would be a good IV. At least not for what we’re looking at. Coming at this from development, I am picky when it comes to IVs, and I’d rather not attempt at 2SLS and be forthcoming about the lack of identification than attempt one with a questionable IV.